Results from our previous studies point to an intimate crosstalk between tumor cells and stromal cells. Recent advances in stem cell research allow us to discuss exciting new hypotheses on tumor progression and etiology in this supplement to the parent grant. 1) Endothelial cells in blood vessels are derived from bone marrow precursors. 2) Fibroblasts in tumors are derived from precursors of the mesenchymal bone marrow stem cell pool. 3) Melanoma cells are derived from melanocyte precursors and not mature melanocytes. We propose to use the supplemental funding for two aims. Aim 1: Determine the mechanisms of fibroblast-dependent endothelial cell differentiation using organotypic cultures. Our preliminary studies clearly demonstrate that fibroblasts are very important for endothelial cell differentiation under in vivo-like conditions. There are apparently two mechanisms: a) Soluble factors secreted by fibroblasts induce endothelial migration and differentiation. b) Differentiation of endothelial cells is dependent on direct cell-cell contact with fibroblasts. We propose that N-cadherin mediates adhesion and allows gap junction formation and we will test this hypothesis. We assume that additional adhesion receptors function as coreceptors, and we will identify them by performing microarray experiments. We also assume that transcriptional regulation of cell surface marker expression is similar to our findings in melanocyte-keratinocyte interactions, and we will explore the importance of major signaling pathways such as PI3 kinase, PKB, PKA and MAPK. From these studies we expect important new information on stromal cells that infiltrate melanomas and we will begin to clarify their biological importance in progression. Aim 2: Differentiate human embryonal stem cells and bone marrow-derived stem cells into melanocyte recursors and melanocytes using organotypic cultures of the human skin. Our strategy is twofold: a) Induce growth factors in fibroblasts and keratinocytes using adenoviral vectors. We will select those growth factors that are important for melanocyte development in the mouse, including PDGF, SCF, ET-3 and bFGF. b) Transduce the stem cells with transcription factors that play important roles in melanocyte differentiation from neural crest cells, including PAX3, MITF and Sox 10. Differentiation of pluripotent stem cells towards a melanocyte phenotype will be verified using antibodies for Tryp-1, Tryp-2 and tyrosinase. We expect to be able to characterize precursor cells for normal melanocytes. Once characterized, we will incorporate these cells into skin reconstructs and graft them onto SCID mice for exposure to UVB and growth factors. If we confirm the hypothesis that melanomas are derived from melanocyte precursors, we can proceed to test the clinical relevance of these findings.